OpenAI's Math Breakthrough List Skips Cryptography, Researcher Notes
thomasahle · x · 2026-10-07
thomasahle is surprised OpenAI's list of AI-produced math results has no cryptography section. While praising results like 'entanglement without secret key' and depth-3 circuit hardness, he notes the absence of new trapdoor functions, LWE hardness, worst-case-to-average-case reductions and more — arguing smarter AI makes provably hard cryptography urgent.
More from Safety
- Uber's Persona Guardrail lifts agent defense accuracy to 95.9%, cuts false approvals to 4.7% — amaarora · 2026-10-07
- Offensive AI Con take: 2025 was 'we need benchmarks', 2026 is 'drowning in exploits' — dyn___ · 2026-10-07
- Multi-tenant RAG leaks via prompts: retrieval-layer ACLs hit 0% leaks over 823,996 canary checks — Acceptable_Army_6472 · 2026-10-07
- Anaconda Combines Coding Agent Swarms and AI Attack Testing After Three Acquisitions — shashib · 2026-10-07
- Germany pushes fast EU approval for Tesla FSD, allowing up to 10% over speed limits — mitchdeg · 2026-10-07
- Aaron Levie: vibe-coded bugs and agentic attacks will redefine enterprise cybersecurity — mattwbaker · 2026-10-07